Detection of any leakage of gaseous hydrogen (GH2) from hydrogen-fueled vehicles is critical in preventing the accumulation of flammable and explosive concentrations of this gas. Concentrations of GH2 greater than or equal to approximately 4% in air are flammable and can be explosive. GH2 leakage detection will also be of great importance for safety in transport and storage in the emerging hydrogen fuel-celled infrastructure. In hydrogen-fueled launch vehicles, during cryogenic storage and subsequent transport of liquid hydrogen, leaks can occur that are associated with the failure of sealed connections at the low temperatures required.
The standard method for the detection of leaks is the use of mass spectrometers. Mass spectrometers have sufficient chemical specificity detection capability to allow detection of leaking amounts of GH2 as opposed to detection of other atmospheric contaminants, such as oxygen, nitrogen, carbon disulfide, carbon monoxide, and methane that exist in an operating environment. However, mass spectrometers have a relatively slow response time in detection of leaking GH2 when applied to large vehicles, such as liquid motor rockets, and are expensive to purchase and to operate.
A widely used alternative in the detection of GH2 is the palladium based sensor approach. Most palladium sensors are based on reversible changes in the physical or electronic properties of palladium in the presence of GH2 or are based on use of palladium as a catalyst for certain reversible chemical reactions in detection of GH2. The limitations of these palladium type sensors arise from the slow time response of the sensor, typically minutes, and the sensitivity to other external environmental factors such as temperature.
It would be useful to provide a GH2 sensor that has sufficient chemical specificity capability, as well as a fast response, and is capable of operating for long periods of time without the need for repairs, recalibration, or replacement. A sensor scheme based upon optical absorption would provide the required sensitivity and specificity. J. Michael Shull (in Astrophysical Journal, vol. 224, pages 841-847, 1978) demonstrated a scheme employing H2 resonance fluorescence with Lyman alpha light (121.567 nm). The major limitation with this approach is that the extreme ultraviolet light employed is strongly attenuated in air, and in the presence of GH2 could even stimulate the explosive chemical reactions that are sought to be avoided, so there can be no other gases present in the test region. This is a significant limitation in the general utility of the invention.
Integrated Cavity Output Spectroscopy (ICOS) methods and associated instruments employ optical absorption cells for spectroscopic purposes. These spectroscopy methods and instruments have a broad range of other applications, such as characterizing mirror reflectivities, determining optical cavity losses (including scattering, absorption, etc.) and measuring thin film absorption. This technique can also be used in the detection of various chemical species. The ICOS method has been used for the detection of trace concentrations of various gas-phase chemical species by measuring the wavelength resonant absorptions that arise from the electronic and vibrational structure of the chemical.
However, the detection of gaseous hydrogen molecules using optical absorption spectroscopy methods such as ICOS has been unsuccessful to date, since there are no easily accessed electronic states to probe and since there is no strong vibrational transition to use as a probe since these are all forbidden for a homonuclear diatomic molecule. Uwe Fink, T. A. Wiggins, and D. H. Rank (Journal of Molecular Spectroscopy, Vol. 18, pages 384-395, 1965) identify specific transition frequencies and absorption line strengths of hydrogen that are extremely weak and not feasible to use as a diagnostic probe using conventional absorption spectroscopy.
Since the hydrogen line widths are broadened due to the Doppler broadening of this light molecule, other sensitive techniques such as frequency modulation (cf., G. C. Bjorklund, M. D. Levenson, W. Lenth, and C. Oritz, “Theory of lineshapes and signal-to-noise analysis”, Appl. Phys. B, vol. 32, page 145 (1983)) are not viable solutions to the problem.